Kindle Notes & Highlights
Read between
October 8 - October 22, 2025
The absolute and relative performance of broad asset classes is systematically related to macroeconomic trends, both cyclical and secular. Among these trends are those of inflation, interest rates, real gross domestic product (GDP), profits, and income growth.
The book will focus on how business cycle patterns relate to investment strategy, such as which sectors do well in which parts of the business cycle, when do equities beat bonds and vice versa, and why. These considerations are key inputs for deciding when to overweight or underweight particular assets.
Globalization has created secular trends, causing rising income inequality in the developed world, while lifting living standards for the world’s poorest people at the fastest rate in human history.
Generally, the best places to look for likely disasters are the sources of cyclical excesses. Cyclical excesses tend to arise in areas in which credit or valuations are expanding unsustainably because of a systemic breakdown in prudential standards for lending and/or investing.
This focus on a snapshot in time outside the broader context often creates a consensus of meaningless information. When payrolls grow by 160,000 in a month when 200,000 jobs were expected, a lot of negative commentary follows, as if a 40,000 miss tells us something about the strength of the labor market. Every month in the United States, several million new jobs are created, and several million old jobs end. A net gain of 160,000 or 200,000 is a very small percentage of that massive churn, and the 40,000 difference is a statistically insignificant piece of information.
In fact, given the wide confidence intervals around the initial estimates, subsequent revisions commonly change that initial interpretation by 180 degrees. By that time, however, the news has moved on to the noise in the next month’s estimate.
Market professionals pontificate in the media as if there is an important kernel of information in the latest noise.
This example illustrates how noise in the data creates volatility in markets. While traders exploit these misinterpretations, longer-term investors need to look through the noise to the underlying signals from the data to make sensible investment decisions.
Assuming monetary policy continues to preempt deflationary forces, inflation has made a round-trip back to subdued 1950s levels. Looking forward, if the Fed remains committed to price stability around 2 percent, inflation would largely be cyclical around a trendless 2-percent rate. This example illustrates the importance of historical and fundamental knowledge about what causes inflation as well as quantitative modeling of trends and cycles.
For example, real GDP responds more quickly to monetary stimulus than inflation, which exhibits more inertia. Given the variability in lags from cycle to cycle and the difference in lags across economic variables, it is not surprising that econometric models based on statistical averages over multiple cycles have a hard time forecasting accurately.
Cycles in interest rates, profits, equity prices, and economic growth are generally related to the underlying business cycle, which we discuss in depth in the next chapter.
Since World War II, for example, expansions have been longer and recessions less frequent than was the case for the business cycles in the prior century. We will discuss why that has been the case in chapter 2. With the time frame of a business cycle in mind, the easiest way to define a cycle is the pattern of movement in an economic or financial variable, such as corporate profits, over the course of an economic expansion and recession.
Similarly, an interest-rate cycle begins when rates rise off the prior cycle’s low point, then peak and fall before rising in the next cycle.
Within a business cycle, which will generally be our reference cycle, there can be multiple cycles in some of the variables we discuss. For this reason, and because it is a good framework for investment strategy, we will usually orient cycles of particular variables over the business cycle defined by the National Bureau of Economic Research (NBER) cycle-dating committee.
Another reason why econometric modeling is limited is that variables switch from trending (or nonstationary) to cyclical (stationary), and the statistical procedures that apply depend on this distinction.
For example, inflation was a nonstationary variable during the period when it trended up and down. It is not a coincidence that econometric tests for stationarity got a lot more attention starting in the 1970s.
macroeconomic theories are more likely to have conceptual and measurement issues, making modeling less reliable.
One example of why this is the case is embodied in Goodhart’s Law, which refers to the notion that “when a measure becomes a policy target, it ceases to be a good measure.”
Put another way, if we get very close to an object in order to describe it, we lose perspective on how it fits into the dynamics of its broader environment.
In fact, some of the best investment opportunities arise when a strategist has confidence in a particular outlook when the market consensus is highly uncertain. Less often, and even more valuable, are cases in which the consensus is strong, but wrong. In this case, a strategist who sees why the consensus is wrong usually has the best opportunities to get into an undervalued asset class early.
Now, the key point is that at any particular point in time there are trillions of data points about the economy. Some are contracting and some are expanding. As more and more data become expansionary, we can better perceive the image of the young woman, while that of the crone fades. Then, as the expansion ages and more economic bits start to contract, the image of the young woman transitions to the aged crone.
Top-down coercion can force that organization, but historical evidence suggests that the long-term structural resilience of the US economy relative to other periodic contenders for its preeminent status probably arises from the more efficient resource allocation created by micro-incentive structures compared to top-down directives that almost by definition rule out efficiency in this framework.
For example, one of the big misconceptions over the first few years of the recovery that began in mid-2009 that caused widespread misinterpretation of economic conditions and missed investment opportunities was a persistent association of declining labor force participation with a weak labor market and worker discouragement over finding jobs. Instead, a deep dive into the issue clearly showed a declining structural trend in labor force participation, mostly as a consequence of an aging population reaching retirement age.
Just as important, the returns to higher education that defer labor force participation to an older age have increased dramatically relative to the returns to less education. Unemployment correlates strongly with education.
In sharp contrast to the highly educated who benefit greatly from globalization, the less educated part of the US population has been thrown into a more intense labor market competition with a vast, similarly undereducated population around the world.
In a nutshell, the modern synthesis of appropriate economic policy boils down to avoiding deflation by keeping long-term inflation expectations stable at a slightly positive level.
Essentially, policy, based largely on the lessons deduced from Keynes’ and Friedman’s analyses of the Great Depression, has stopped the frequent deflationary collapses that characterized the US economy prior to World War II.
The absence of deflation in the modern era has dramatically reduced the amplitude of business cycles since 1950 (Exhibit 2.1). Because inflation has generally stayed positive, it has been rare to see a year of negative nominal GDP growth. In fact, the financial crisis of 2008–2009 and the associated recession was the first instance of negative nominal GDP growth since the 1940s and the worst since the 1930s, helping to explain the unprecedented and highly controversial policy response at the time.
Its insane that an evolution of monetary policy can weed out such severe bugs. But how long is a zero deflation policy possible. Does the economy not build scaffolding on top of an artificially stable business cycle?
While eliminating deflation has played a key role in moderating business cycles, there are several additional factors that seem to have damped the amplitude of modern business cycles. Information technology has made real-time inventory management and more efficient supply chain management possible, largely eliminating one of the major sources of old-fashioned business cycles: much bigger under- and overshoots of inventory accumulation.
As we shall see, goods-producing, and especially durable-goods-producing industries tend to be much more cyclical than modern services industries, such as health care, education, and government. A bigger share of jobs in less cyclical industries helps smooth economic growth compared to an economy in which manufacturing jobs dominate.
In fact, there is also a case to be made that the policy safety net has increased moral hazard in a way that allows reckless behavior to accumulate over cycles rather than purging it in each cycle as was the case in “the good old days” of more frequent recessions and depressions.
Leverage is a one-way ticket to higher wealth if the leveraged asset only goes up in price. The housing collapse was a useful reminder that leverage works both ways. As confidence about the house price outlook waxed and waned, strong cyclical forces were unleashed in both directions.
Variables of interest for analyzing business cycles include employment, unemployment, household incomes, retail sales, spending on durable goods like automobiles, housing investment, industrial production, business orders, shipments, inventories, commercial construction, profits, revenues, consumer inflation, producer price inflation, wage inflation, productivity, unit labor costs, and GDP inflation, to name some of the main categories. In addition, financial variables, like interest rates, credit, and money supply are cyclical. We will discuss the business-cycle behavior of financial
In addition to more moderate cycles since 1950, it is also a fact that expansions have been much longer and recessions much shorter on average in the modern era. The economy has spent a much larger proportion of time growing in the era of activist policy.
These four components are employees on nonfarm payrolls, real personal income less transfer payments from the government, industrial production, and real manufacturing trade and sales. Basically, a recession involves a suitably prolonged decline in production, jobs, incomes, and spending.
Interest rate spread (yield curve spread), ten-year Treasury note yield minus Federal funds rate. An inversion of this spread (lower ten-year yield than the Fed funds rate) has proved one of the most reliable precursors of recessions.
The stock market itself is a leading indicator because it anticipates the twists and turns in all the different industries that make up the global economy.
Industrial production, a coincident indicator, usually reaches its peak right as a recession (shaded areas) begins (Exhibit 2.3). Employment, another coincident measure of aggregate economic activity also tends to peak when recessions begin. For this reason, it is often the case that people feel best just when an expansion is ending. Conversely, they tend to feel the worst when a recession is ending because economic activity, measured by coincident indicators like industrial production and employment, for example, tends to be at its cyclical low point just as a recovery begins.
Over the course of the business cycle, confidence about the need for new investment fluctuates with the strength of the economy.
One reason recessions end is this caution about spending starts to appear excessive relative to the business opportunities still available or becoming apparent.
Inventory excesses were a bigger force driving the business cycle before sophisticated information-processing and logistics technology existed. Nevertheless, inventory accumulation in the capital goods sector remains much more volatile (cyclical) than the overall economy (GDP).
Volatility in household durables spending is consequentially much greater than spending on everyday necessities. That’s why consumer staple stocks are considered defensive plays for bad times, while consumer discretionary companies thrive in better economic environments.
It also illustrates the point that “what’s different” about a particular business cycle is often a function of where excessive financing has been applied. In this case, it was residential real estate. In the earlier 1992–2001 cycle, rather than excessive credit creation, it was excessive equity valuations that funded an investment boom. While the excessive equity valuations did set the stage for a secular bear market, the recession that followed the technology bust was relatively mild and short lived, unlike the 2007–2009 recession.
The difference between these two cycles also illustrates the point that cycles funded more by equity valuation excesses are generally less damaging than those fueled by excessive debt growth. We will discuss this difference in more detail in chapter 8.
As a mix of cyclical and less cyclical jobs, overall employment is one of the best trackers of whether GDP is growing or not. That’s why it is one of the four components in the index of coincident indicators. These four components are available on a monthly basis, making it possible to date recessions in the month they start and end.
Conversely, right before a recession, when current conditions are the best, consumers recognize that the economy does not have as much room to improve in the coming months and the gap tends to be most negative (future expectations are below current conditions measures).
Over the course of the business cycle, monetary policy goes through a cycle from maximally accommodative to maximally restrictive. Generally, monetary policy is most stimulative in the first phase of expansion, when inflation is low and unemployment high. This is reflected in a very steep yield curve that embodies cycle lows in short-term money market rates along with higher rates in longer-maturity bonds that recognize that these short rates are temporary until the economy absorbs more of the slack in labor markets and production capacity.
As the recession worsens and the inflation rate falls, monetary policy usually moves rapidly to cut short-term rates, and the cycle begins anew.
Aside from labor market slack, there is slack in industrial capacity that varies over the business cycle. As can be seen in Exhibit 2.11, the capacity utilization rate in manufacturing tends to bottom at the end recessions with maximum slack available for growth.
What is capacity utilization and how is it quantified?
Automation limits the impact of higher wage costs in manufacturing as factories experience the loss of jobs that farms experienced over one hundred years ago.
